IMMIGRATION STRESS AND RELATIONSHIP SATISFACTION IN LATINO COUPLES: THE ROLE OF DYADIC COPING

Transcription

1 IMMIGRATION STRESS FALCONIER et al. Journal of Social and Clinical Psychology, Vol. 32, No. 8, 2013, pp IMMIGRATION STRESS AND RELATIONSHIP SATISFACTION IN LATINO COUPLES: THE ROLE OF DYADIC COPING MARIANA K. FALCONIER Virginia Polytechnic Institute and State University FRIDTJOF NUSSBECK University of Bielefeld, Germany GUY BODENMANN University of Zürich This study examined (a) the relationship of Latino partners overall immigration stress and each of its different dimensions with their relationship satisfaction and (b) whether a partner s support (supportive dyadic coping) and the couple s conjoint efforts to cope with stress (common dyadic coping) can moderate those relationships. An Actor-Partner Interdependence Model (APIM; Kenny, 1996) was used to analyze self-report data collected from 104 Latino immigrant couples in the U.S. Structural equation modeling results suggest that common dyadic coping and the supportive dyadic coping provided by the male partner can attenuate the negative association of various aspects of Latinas immigration stress mostly with her relationship satisfaction and to some extent with her male partner s as well. By contrast, common dyadic coping and the female partner s supportive dyadic coping do not play any moderating role in the relationship between most dimensions of the male immigration stress and either partner s relationship satisfaction. Limitations, research, and clinical implications are discussed. According to the 2010 U.S. Census data Latinos have become the first minority in the country with a population of 50.5 million peo- Address correspondence to Mariana K. Falconier, PhD, Virginia Polytechnic Institute and State University, 7054 Haycock Rd, Suite 202C, Falls Church, VA 22043; marianak@vt.edu 2013 Guilford Publications, Inc. 813

2 814 FALCONIER ET AL. ple (Humes, Jones, & Ramirez, 2011). First generation immigrants, who still constitute a large proportion (37.4%) of Latinos living in the U.S. (U.S. Census Bureau, 2009), are likely to cope with significant levels of stress associated with the demands of immigration. Latino immigrants as well as other immigrant groups experience stress when dealing with discrimination, material, and emotional losses, not feeling at home, missing the family left in their countries, occupational challenges, difficulties accessing services in the new country, and, obviously, cultural differences including language barriers (Aroian, Norris, Tran, & Schapper-Morris, 1998; Goodkind, Gonzáles, Malcoe, & Espinosa, 2008). The research on the effects of immigration related stressors on Latinos has focused almost exclusively on the effects of acculturative stress on the individual. This line of research has linked acculturative stress to poor physical health (Caplan, 2007; Finch & Vega, 2003) and various negative psychological outcomes such as depression, anxiety, drinking, substance abuse, and eating disorders (Caetano, Ramisetty-Mikler, Caetano Vaeth, & Harris, 2007; Caplan, 2007; Revollo, Qureshi, Collazos, Valero, & Casas, 2011; Sarmiento & Cardemil, 2009; Torres, 2010). Despite its contributions, the exclusive focus on acculturative stress, which refers to the stress generated by the problems resulting from conflicts between the immigrant and the host cultures (Berry, 2006), has left out other immigration related stressors (e.g., missing family) that do not stem from the contact with a different culture. The few studies that have actually focused on the stress associated with various aspects of the immigration experience (e.g., Aroian, Kaskiri, & Templin, 2008; Aroian, Norris, González de Chávez Fernández, & García Averasturi, 2008), instead of examining acculturative stress only, suggest that factors such as not feeling at home, a sense of loss, discrimination, language and occupational challenges, and novelty can have different associations with individual psychological distress. However, given that contextual factors (e.g., cultural similarities and differences between the immigrant group and the host country or geographical distance between the country of origin and the receiving country) may contribute to exacerbate certain dimensions of immigration stress, results have been inconsistent across studies. For example, one study (Aroian, Kaskiri, et al., 2008) with Arab immigrant women related depression with all immigration stress dimensions except for occupational challenges. Differently, another study linked psychological distress with novel-

3 IMMIGRATION STRESS 815 ty in both men and women but with loss, occupation, and language difficulties only in women and not feeling at home only in men (Aorian, Norris, et al., 2008). Interestingly, even though this is the only study that has examined different dimensions of immigration stress in a Latino immigrant population, these immigrants were living in the Canary Islands, Spain, a country whose similarities in values and language to Latin American countries limits the possibilities of extending the findings of that study to Latino immigrants living in the U.S. In addition to the absence of research on the various dimensions of immigration stress in the Latino immigrant population in the U.S., most important perhaps it is the extent to which the focus on the individual in the acculturation stress studies and in the immigration stress research has ignored the potential effects of immigration stress on Latino couples relationships. With the exception of two studies, which found a positive association of acculturative stress with interpartner violence (Caetano et al., 2007) and decreased relationship satisfaction (Negy, Hammon, Reig-Ferrer, & Carper, 2010), there has been, to our knowledge, no other study published in English on Latino or other immigrant populations that examined the extent to which couples functioning may be affected by immigration stress. And even these two studies were limited in that they both focused only on acculturative stress and they either failed to incorporate both actor and partner effects (Negy et al., 2010) or to just focus on immigrant couples (Caetano et al., 2007). As noted by Padilla and Borrero (2006) the ways in which Hispanics can be resilient to the negative consequences of stress in their lives generally, and more specifically in their marital relationships is in need of investigation (p. 301). The relative lack of research on the effects of immigration stress on couples relationships contrasts with the couples stress literature that has long conceptualized and documented the adverse effects of one partner s stress on both the relationship and the other partner s emotional well-being (for an overview see Randall & Bodenmann, 2009; Story & Bradbury, 2004). According to the systemictransactional model (Bodenmann, 2005), a model for understanding coping processes in the context of intimate relationships, one cannot examine one partner s stress appraisals or coping efforts without considering the effects on the other partner and the marriage (Bodenmann, 2005, p. 36). Findings from studies on breast cancer (Badr, Carmack, Kashy, Cristofanilli, & Revenson, 2010; Kay-

4 816 FALCONIER ET AL. ser, Watson, & Andrade, 2007), economic problems (Falconier & Epstein, 2010), and chronic daily stressors (Bodenmann, Ledermann, & Bradbury, 2007) indicate that each partner s experience of stress contributes to the other partner s stress and declines in relationship satisfaction (for a review see Randall & Bodenmann, 2009). When both partners experience similar stress as it might be the case with immigrant Latino couples, it is possible that each partner s satisfaction with their relationship may be even more negatively affected. Findings from the couples stress research literature and the two studies that have linked acculturative stress in Latinos with interpartner violence and lower marital satisfaction in women suggest that immigration stress may significantly contribute to declines in relationship satisfaction in immigrant Latino couples. The present study is the first to examine the association of overall immigration stress and its different aspects (missing family, occupational and language challenges, discrimination, novelty, sense of loss, not feeling at home) with relationship satisfaction in couples applying an Actor-Partner Interdependence Model (APIM; Kenny, 1996). Such an approach allows for the examination of both actor and partner effects and the control for partners interdependence. Considering the potential negative effects of overall immigration stress and each of its different aspects on couples relationships, this study also seeks to examine whether dyadic coping can protect partners relationship satisfaction from such effects. According to the systemic-transactional model (Bodenmann, 2005), dyadic coping refers to the positive and negative strategies that couples use to cope with stress together as a unit. Positive strategies include supportive, delegated, and common dyadic coping. Supportive Dyadic Coping (SDC) describes one partner s efforts to assist the other partner cope with his or her stress by providing emotional (e.g., empathic understanding) and/or problem-focused support (e.g., practical advice). Delegated Dyadic Coping (DDC) involves a partner s attempts to assist the other partner by taking on his or her tasks and duties. Common Dyadic Coping (CDD) describes partners conjoint efforts to cope with stress by engaging in joint problem solving, joint information seeking, sharing of feelings, mutual commitment, or relaxing together (Bodenmann, 2005, p. 38). Negative dyadic coping refers to partners strategies that can be characterized as hostile (minimizing, withdrawing, blaming), ambivalent (helping unwillingly or showing it is not necessary), and/or superficial (helping insincerely). Only supportive and common dyadic cop-

5 IMMIGRATION STRESS 817 ing have been found consistently to be associated with relationship satisfaction and reduced verbal aggression (Bodenmann, Meuwly, Bradbury, Gmelch, & Ledermann, 2010; Papp & Witt, 2010). Despite these positive associations and even though from a systemic-transactional perspective positive dyadic coping should attenuate the effects of stress on couple s functioning, only two studies have actually examined this possibility. One of these studies (Bodenmann, Atkins, Shär, & Poffet, 2010) used data from female college students and did not find overall dyadic coping to moderate the relation between the partners daily hassles and their sexual activity. Nonetheless, the other study (Bodenmann, Meuwly, et al., 2010), which also collected data from only one partner, found that supportive and common dyadic coping did moderate the association between daily stress and verbal aggression, adding to the evidence that those two forms of dyadic coping may be particularly beneficial to couples functioning. It is possible that supportive and common dyadic coping can serve to attenuate the possible negative link between each Latino partner s immigration stress and their relationship satisfaction. Latinos strong family orientation, commonly referred to as familismo (Lugo Steidel & Contreras, 2003), values family interdependence and loyalty and views the family as a source of emotional and instrumental support. Providing support to a stressed partner and engaging in conjoint coping efforts when both partners are experiencing stress seems to be consistent with Latinos familismo. Thus, we would expect common dyadic coping to be particularly important in Latino couples. In brief, the purpose of the present study with Latino couples is to examine (a) whether and to what extent overall immigration stress is associated with each partner s relationship satisfaction, (b) whether this association exists for each aspect of immigrations stress (missing family, novelty, discrimination, not feeling at home, sense of loss, language barrier, occupational challenge), and (c) whether supportive and common dyadic coping moderate the relationship of overall immigration stress and each aspect of immigration stress with each partner s relationship satisfaction. Given that the literature has reported all of the dimensions of immigration stress considered in this study to affect both male and female Latino immigrants and other immigrant groups (e.g., Aorian, Norris, et al., 2008; Caetano et al., 2007; Goodkind et al., 2008; Revollo et al., 2011) and that results have been inconsistent in previous studies (e.g., Aroian,

6 818 FALCONIER ET AL. FIGURE 1. Conceptual Model 1. FIS = Female Immigration Stress; MIS = Male Immigration Stress; MSDC = Male Supportive Dyadic Coping; FSDC = Female Supportive Dyadic Coping; FRS = Female Relationship Satisfaction; MRS = Male Relationship Satisfaction. The covariance between MSDC and FSDC were not depicted for clarity of presentation. Kaskiri, et al., 2008; Aroian, Norris, et al., 2008), no specific hypotheses are formulated for each immigration stress dimension. Instead, this study adopts an exploratory approach in order to understand whether only some immigration stress dimensions are related to relationship satisfaction and dyadic coping. Due to sample size restrictions, this study examined two models, one for supportive dyadic coping (Figure 1) and another one for common dyadic coping (Figure 2) testing the following hypotheses: 1. Each partner s immigration stress (overall and by type) will be negatively associated with one s own (actor effect) and the partner s (partner effect) relationship satisfaction (Figures 1 and 2). 2. The association of each partner s immigration stress (overall and by type) with one s own and the partner s relationship satisfaction will be moderated by the other partner s supportive dyadic coping. The partner s supportive dyadic coping will attenuate the negative association between immigration stress (overall and by type) and relationship satisfaction (Figure 1).

7 IMMIGRATION STRESS 819 FIGURE 2. Conceptual Model 2. FIS = Female Immigration Stress; MIS = Male Immigration Stress; CDC = Common Dyadic Coping; FRS = Female Relationship Satisfaction; MRS = Male Relationship Satisfaction. 3. The association of each partner s immigration stress (overall and by type) with one s own and the partner s relationship satisfaction will be moderated by the couple s common dyadic coping. Common dyadic coping will attenuate the negative association between immigration stress and relationship satisfaction (Figure 2). 4. There will be positive, direct associations in immigration stress (overall and by type), supportive dyadic coping, and relationship satisfaction between partners (Figures 1 and 2). METHOD SAMPLE The sample consisted of 107 immigrant Latino couples residing in the Washington, D.C. metropolitan area. These community couples were recruited through flyers and presentations in churches, festivals, medical, and therapists offices, and community agencies. Inclusion criteria were self-identification as Latino/a, being at least 18

8 820 FALCONIER ET AL. years old, and having lived with the current partner for at least one year. Given the focus on immigrant couples in the present study, seven of the original 114 couples had to be excluded since at least one of the partners was not foreign born or had immigrated to the U.S. before the age of 10. On average men were years old (SD = 8.68) and women were years old (SD = 7.7). The mean length of residence in the U.S. was years (SD = 8.03; range 1 to 35 years) for men and years (SD = 7.43; range 1 to 32 years) for women. Participants had emigrated from various countries including El Salvador (32.7%), Peru (13.9%), Mexico (12%), Guatemala (11.1%), and other Latin American countries (30.3%). Even though the sample is not representative of the Latino population in the U.S. (64.5% Mexican, 3.2% Salvadoran, 2.0% Guatemalan, 1.0% Peruvian; U.S. Census Bureau, 2009), it is representative of the Latino population living in the Washington D.C. metropolitan area (e.g., Montgomery County: 9.4% Mexican, 32.2% Salvadoran, 7.2% Guatemalan, 6.3% Peruvian; U.S. Census Bureau, 2009). The educational level ranged from 27.9% of men and 30.7% of women not having completed high school to 21.2% of men and 24% of women having graduated at least from college. Most men (84.6%) and women (61.5%) were employed, and partners combined income was over $50,000 for 47.1% of couples, between $20,000 and $49,999 for 31.7%, and below $19,999 for 14.4%. Thirty-two percent of the couples had at least one partner reporting the presence of another household member contributing to the household income. The majority of both men (72.1%) and women (80.8%) were Catholic, whereas only 15.3% of men and 14.4% of women were affiliated to other Christian religions, 1.9% of men and 1 % of women reported to be Jewish, and 7.7% of men and 3.8% of women did not have any religious affiliation. Twenty-eight couples (26.9%) were cohabiting whereas 76 couples (73.1%) were married. On average couples had been living together for years (SD = 7.66) and had 1.84 (SD = 1.22) children younger than 21 years old living in the household. PROCEDURE Couples interested in participating in the study received a package with consent forms, instructions, and a set of self-report questionnaires. Partners were instructed to sign the consent form, complete

9 IMMIGRATION STRESS 821 their individual assessments separately and without consulting with each other, and return all the materials back by mail or in person. Once both partners signed consent forms and questionnaires were received by the researchers, a $25 gift card was mailed to each partner. Participants were assigned a three-digit identification number and all identifying contact information was kept confidential and not included in the database. INSTRUMENTS Participants completed a set of self-report questionnaires and a demographic information form. Except for the demographic form that was developed in Spanish, all of the assessment instruments were originally in English. These instruments were first translated into Spanish by a native speaker and then back-translated into English so that both English versions could be compared and the translation could be adjusted accordingly. Immigration Stress. The Demands of Immigration Scale (DIS; Aroian et al., 1998) was used to measure stress associated with the immigrant status. In the present study, respondents had to rate the frequency with which they experienced 23 different immigration stress events on a three-point Likert-type scale: 0 = never, 1 = sometimes, 2 = very often. The DIS has six different subscales, each of them corresponding to a different dimension of immigration stress: Loss (e.g., I feel sad when I think of special places back home), Novelty (e.g., I am always facing new situations and circumstances), Occupation (e.g., The work credentials I had in my original country are not accepted), Language (e.g., I have difficulty doing ordinary things because of a language barrier), Discrimination (e.g., As an immigrant, I am treated as a second-class citizen), and Not at Home (e.g., I do not feel that this is my true home). Considering Latinos strong family orientation and the fact that the DIS did not include any specific item referring to the family left in the country of origin, a Missing Family subscale was added. This subscale included one of the items from the Loss subscale (I miss the people I left behind in my original country) and four new items (e.g., I wish I could speak with my family in my country more often). Confirmatory Factor Analysis showed that the newly formulated items load on a single scale. The internal consistency for the new scale was.73 for men and

10 822 FALCONIER ET AL..76 for women. The internal consistency for the total scale including the Missing Family subscale was.93 for both men and women (all other consistencies can be found in Table 1). The DIS has been used in various studies of immigration with the total score as an indicator of overall immigration stress and the score of each subscale as an indicator of each immigration stress dimension (for a review see Aroian et al., 2008) and has been reported to have a total internal consistency ranging from.91 to.94 with a test-retest reliability after three weeks ranging from.89 to.92 (Aroian et al., 1998). Dyadic Coping. Supportive and common dyadic coping were assessed using the Dyadic Coping Inventory (DCI; Bodenmann, 2008). The DCI is a 37-item self-report questionnaire with ten different subscales: Stress Communication by oneself and by partner, Supportive Dyadic Coping by oneself and by partner, Delegated Dyadic Coping by oneself and by partner, Negative Dyadic Coping by Oneself and by partner, Common Dyadic Coping, and Overall Evaluation of Dyadic Coping. All items are rated on a five-point Likert-type scale ranging from 1 = very rarely to 5 = very often. The DCI has been used extensively in various studies (for a review see Bodenmann, 2008) and the convergent and factorial structure of its subscales have been confirmed in Swiss, German, French, Italian, and Latino population (Bodenmann, 2008; Donato et al., 2009; Falconier, Nussbeck, & Bodenmann, 2013; Ledermann et al., 2010). In the Supportive Dyadic Coping subscale, respondents are asked to rate their partner s tendency to provide both problem-focused (e.g., analyzing the situation) and emotion-focused (e.g., showing empathy, listening) support. In this study the internal consistency was.89 for both, males and females. In the Common Dyadic Coping subscale partners are asked about the couple s joint efforts to deal with stressors affecting both partners. Three of the items describe emotion-focused coping strategies (e.g., relaxing together) whereas the other two items involve problem-focused coping efforts (e.g., searching for solutions together). Both partners responses were averaged so that both of their perceptions of common coping could be captured. The t-tests for paired samples did not indicate any statistically-significant differences between males and females perceptions of common coping, MD =.34, t (91) =.68, p =.50. The internal consistency of this subscale was.94 for men and.92 for women. Relationship Satisfaction. The Dyadic Satisfaction subscale from the Dyadic Adjustment Scale (DAS; Spanier, 1976) was used to measure

11 IMMIGRATION STRESS 823 TABLE 1. Means, Standard Deviations, T-tests, and Internal Consistency (N = 104 females and 104 males) Variable Mean SD Mean Difference SD (Mean Diff.) T P α Overall Immigration Stress Females Males Immigration Stress Dimensions Missing Family Females Males Novelty Females Males Not at Home Females Males Language Females Males Occupation Females Males Discrimination Females Males Loss Females Males Common Dyadic Coping Supportive Dyadic Coping Females Males Relationship Satisfaction Females Males

12 824 FALCONIER ET AL. relationship satisfaction. This subscale includes 10 Likert-type items measuring the respondents satisfaction with their relationship (e.g., How often do you think that things between you and your partner are going well?). The other three subscales from the DAS were excluded as they have been found to confound the measurement of relationship satisfaction with behaviors that contribute to satisfaction (Kurdek, 1992). The DAS is an instrument that has been used widely in couples research (for a review see Graham, Liu, & Jeziorski, 2006). In the present sample the internal consistency of the dyadic satisfaction subscale was.86 for both men and women. ANALYTIC STRATEGY T-tests for paired samples were conducted to assess gender differences on immigration stress (overall and by subscale), supportive and common dyadic coping as well as relationship satisfaction. In principle, dyadic data can be analyzed relying on Multilevel (ML) models or Structural Equation Models (SEM) or a combination of both. In this contribution, we use the SEM approach because this approach allows for examining relationships between variables free from measurement error, testing the goodness of fit of the base models (see below) and testing the measurement structure of all study variables simultaneously. First, two base models, one for overall immigration stress and one for all the different dimensions of immigration stress, were tested. In these models, no assumptions were made about the structural model part; instead all factors were allowed to correlate. These models served to identify, if the measurement structure concerning the test-halves fit to the data (see below). Factors in these two models included immigration stress (overall or subscales), supportive and common dyadic coping and relationship satisfaction. Factor loadings were restricted to 1 in these and all remaining models for reasons of parsimony. Then, the base models were incorporated into the two structural models depicted in Figures 1 and 2, one for common coping and one for supportive coping. Model fit for all models was assessed by the chi-square (χ 2 ) statistic and given the χ 2 -test s sensitivity to sample size the recommended ratio of no more than 3:1 for the χ 2 to df ratio was used (Kline, 1998). Additionally, the three fit indices recommended by Hu and Bentler (1999)

13 IMMIGRATION STRESS 825 were used to evaluate model fit: Bentler s Comparative Fit Index (CFI), the Standardized Root Mean Square Residual (SRMR), and the Root Mean Square Residual of Approximation (RMSEA). The following cut-off values were used to assess acceptable model fit: CFI >.90, SRMR <.08, and RMSEA <.08 (Hu & Bentler, 1999). We will report 95% confidence intervals (95% C.I.) for the bootstrapped model parameters. Setting the C.I. to 95% allows for an error rate of 2.5% at each side of the distribution. All analyses were conducted using Mplus 6.11 (Muthén & Muthén, ). Base Models. The immigration stress, supportive and common dyadic coping, and relationship satisfaction scales were split into two homogeneous test-halves in order to obtain two indicators for each of the scale (for a discussion of creating test-halves as indicators of latent variables see Bandalos, 2002). This procedure involved the following steps: First, a confirmatory factor analysis (CFA) was run for each latent construct, including each of the immigration stress dimensions, for men and women separately and including each of the scale items as the observed variables; second, those items were rank-ordered with respect to their factor loadings; third, the 1st item was assigned to test-half 1; the 2nd and 3rd items to test-half 2; the 4th and 5th item to test-half 1 again and so on. In most cases the rank-ordering for the item was stable across gender. In the few cases where the rank order was not stable, the differences in the loading coefficients between the items with differing rank ordering were marginal and therefore, these items were assigned arbitrarily to test-half 1 or 2. Nonetheless, the item assignment to the test halves was identical across gender. Structural Models. In order to examine whether supportive and common dyadic coping played a moderating role between immigration stress and relationship satisfaction, latent interaction effects between latent variables were specified using the LMS-approach implemented in Mplus (in order to reduce the computational complexity integration = montecarlo was specified). Goodness-of-fit coefficients cannot be reported for the latent interaction models since Mplus does not provide these. As noted earlier, due to sample size limitations, for overall immigration stress and each immigration stress dimension, two structural models were tested, one with the couple s common dyadic coping and another one with each partner s supportive dyadic coping as the moderating variables.

14 826 FALCONIER ET AL. RESULTS DESCRIPTIVE STATISTICS AND T-TESTS The relationship of demographic variables (length of stay, number of children, age, income, education level, and length of relationship) with immigration stress, supportive and common dyadic coping, and relationship satisfaction were assessed through Pearson bivariate correlations. Results indicated that only length of stay in the U.S. is negatively and modestly associated with immigration stress for both men (r = -.34) and women (r = -.29). T-tests for paired samples did not indicate statistically-significant differences between the partners for almost any of the variables of interest in this study (see Table 1). The only significant gender differences were found in the IMS Loss subscale and in the DCI Supportive Dyadic Coping. Women reported significantly higher levels of loss than their partners did (MD =.12, t = 2.03, p =.05) and men reported higher levels of supportive dyadic coping by their partners than women did (MD =.29, t = 2.63, p =.01). However, the gender difference in the IMS Loss subscale disappeared when length of stay was controlled for. This is consistent with the results of the t-tests for paired samples that indicated that on average men in this sample had stayed significantly more years in the U.S. than women (MD = 1.97 years, t = 2.945, p =.004). BASE MODELS As noted earlier, two base models with correlated latent variables for both partners were tested. The first model that included overall immigration stress, supportive dyadic coping, common dyadic coping, and relationship satisfaction fit the data acceptably: χ 2 (63) = , p <.01; χ 2 /df = 1.70; CFI =.96; SRMR =.10; RMSEA =.08; C.I.RMSEA (90%) =.05,.11. Even though the SRMR did not reach the cut-off value of.08, the χ 2 /df ratio and the CFI and RMSEA values are indicative of acceptable model fit. In the second model instead of the overall immigration stress, the different dimensions of immigration stress were included: missing family, novelty, not at home, language, occupation, discrimination, and loss. This second model fit the data acceptably: χ 2 = ; df = 513; p <.01; χ 2 /df = 1.46; CFI =.90; SRMR =.07; RMSEA =.07; C.I.RMSEA (90%) =.06.08).

15 IMMIGRATION STRESS 827 TABLE 2. Correlations Among Immigration Total Scale and Subscales, Supportive Dyadic Coping, Common Dyadic Coping, and Relationship Satisfaction for Men and Women (N = 104 females and 104 males) 1. F Immigration Stress - 2. M Immigration Stress F Missing Family M Missing Family F Novelty M Novelty F Not at Home M Not at Home F Language M Language F Occupation M Occupation F Discrimination M Discrimination F Loss M Loss MSDC FSDC CDC FRS MRS Note. F = Female; M = Male; FSDC = Female Supportive Dyadic Coping; MSDC = Male Supportive Dyadic Coping; CDC = Common Dyadic Coping; FRS: Female Relationship Satisfaction; MRS = Male Relationship Satisfaction. Coefficients printed in bold type differ significantly from 0 (95% C.I.).

16 828 FALCONIER ET AL. TABLE 3. Structural Models for Overall Immigration Stress with Common Dyadic Coping and Supportive Dyadic Coping: Unstandardized Regression Coefficients, Variances of Dependent Variables and Regression Residuals, and Correlation of the Regression Residuals (N = 104 females and 104 males) Common Dyadic Coping FIS MIS CDC FIS CDC MIS CDC Res. Var. Var. Res. Corr. FRS.68 a.16 p.55 m MRS.11 p.46 a.57 m Supportive Dyadic Coping FIS MIS FSDC MSDC FIS MSDC MIS FSDC Res. Var. Var. FRS.63 a.05 p.10 a.45 p MRS.12 p.25 a.07 p.36 a Res. Corr. Note. FRS = Female relationship satisfaction; MRS = Male relationship satisfaction; FIS = Female immigration stress; MIS= Male immigration stress; CDC = Common dyadic coping. FIS/MIS CDC = interaction term; FSDC = Female supportive dyadic coping; MSDC = Male supportive dyadic coping. FIS/MIS FSDC/MSDC = interaction terms; Res. Var. = Residual (unexplained) variance of the dependent variable; Var. = Variance (unconditional) of the dependent variable; Res. Cor. = Correlation of the residual terms of the dependent variables. a actor effects; p partner effects m main effects; Only unstandardized regression coefficients are presented since there is no statistical theory allowing for standardized regression coefficients with the LMS approach. ( ) indicates models with convergence problems, so results are presented without interaction terms. Parameters printed in bold type differ significantly from 0 (95% C.I.). Table 2 presents the correlations between the latent variables in the unrestricted models (one for overall immigration stress, one for its subscales). As hypothesized, positive correlations across partners were found for relationship satisfaction (.88), supportive dyadic coping (.83), overall immigration stress (.45), and stress associated with missing family (.41), novelty (.52), not feeling at home (.46), language (.60), occupation (.55), and discrimination (.29). Common dyadic coping was positively correlated with both the male (.71) and the female (.65) supportive dyadic coping. Correlations among subscales were all positive and significant for both genders, falling into the range of.42 <= r <=.86 for males and the range of.56 <= r <=.93 for females. STRUCTURAL MODEL Overall Immigration Stress and Common Dyadic Coping. The first structural model that was examined included common dyadic coping as a moderator in the prediction of each partner s relationship satisfaction by each partner s immigrant stress. As depicted in the

17 IMMIGRATION STRESS 829 FIGURE 3. Interactive Effect of Immigration Stress and Common Dyadic Coping on Relationship Satisfaction for Females. For clarity of presentation the partner's immigration stress was set to 0 (mean of the latent variable in the structural models). low CDC = low common dyadic coping (CDC = 1); high CDC = high common dyadic coping (CDC = 1). upper part of Table 3, female relationship satisfaction is statistically predicted by female overall immigration stress (FIS; actor effect, annotated a), male overall immigration stress (MIS, partner effect; p), common dyadic coping (CDC; main effect, m), and the interaction between female (FIS CDC) or male (MIS CDC) overall immigration stress and common dyadic coping. The last three columns present the residual variance of female relationship satisfaction, the total variance of female relationship satisfaction and the correlation of the residuals for female and male relationship satisfaction residuals. Male relationship satisfaction is statistically predicted by the same set of variables, however, the effect of female overall immigration stress is a partner effect (p) and male overall immigration stress is an actor effect (a). As hypothesized (Figure 3), the female relationship satisfaction was negatively associated with her own overall immigration stress (total score) (-.68) but not with the male overall immigration stress (.16). The female relationship satisfaction was also positively related to common dyadic coping (.55). The association between women s own immigration stress and their relationship satisfaction was moderated by the couple s common dyadic coping (.75): Rela-

18 830 FALCONIER ET AL. FIGURE 4. Interactive Effect of Immigration Stress and Common Dyadic Coping on Relationship Satisfaction for Males. For clarity of presentation the partner s immigration stress was set to 0 (mean of the latent variable in the structural models). Low CDC = low common dyadic coping (CDC = -1); high CDC = high common dyadic coping (CDC = 1). tionship satisfaction of female partners in couples with good common coping is less strongly negatively associated to their overall immigration stress (and may even become positively associated if the scores on common dyadic coping are higher than 4.5). There is no moderating effect of common dyadic coping on the association between the male overall immigration stress and the female relationship satisfaction (Figure 4). In contrast to female partners, only one significant association can be found for male partners. Their relationship satisfaction is positively associated with the couple s common dyadic coping (.57; Figure 4). Dividing the residual variances by the total variances of the female and male relationship satisfaction, respectively, provides an estimate of the unexplained variance in the statistical model. The inverse yields the proportion of explained variance. Around 72% of the latent variance of female relationship satisfaction can be explained by the association with overall immigrant stress and common dyadic coping. For male partners, around 53% of the variance of relationship satisfaction can be explained. The residual correlations of relationship satisfaction of r =.18 shows that both partners relationship satisfactions are associated beyond the association with the predictor variables in the model.

19 IMMIGRATION STRESS 831 TABLE 4. Structural Models for Immigration Stress Dimensions with Common Dyadic Coping: Unstandardized Regression Coefficients, Variances of Dependent Variables, and Regression Residuals, and Correlation of the Regression Residuals (N = 104 females and 104 males) FIS MIS CDC FIS CDC MIS CDC Res. Var. Var. Res. Corr. Missing Family FRS MRS Novelty FRS MRS Not at home FRS MRS Language FRS MRS Occupation FRS MRS Discrimination FRS MRS Loss FRS MRS Note. FRS = Female relationship satisfaction; MRS = Male relationship satisfaction; FIS = Female immigration stress; MIS= Male immigration stress; CDC = Common dyadic coping. FIS/MIS CDC = interaction term; Res. Var. = Residual (unexplained) variance of the dependent variable; Var = Variance (unconditional) of the dependent variable; Res. Cor. = Correlation of the residual terms of the dependent variables. Only unstandardized regression coefficients are presented since there is no statistical theory allowing for standardized regression coefficients with the LMS approach. Parameters printed in bold type differ significantly from 0 (95% C.I.).

20 832 FALCONIER ET AL. Overall Immigration Stress and Supportive Dyadic Coping. The second part of Table 3 presents the associations between relationship satisfaction, overall immigration stress and supportive dyadic coping. Since both partners can provide supportive dyadic coping, also two variables have to be considered. Supportive dyadic coping by one partner was analyzed to have a moderating impact on the association between the other partner s immigration stress and each partner s relationship satisfaction. Estimation problems, which were not resolved even after additional iterations, prevented us from arriving to final results in the models that included overall immigration stress. Even though the interaction effect could not be estimated, in models without the interaction term the female relationship satisfaction was significantly (-.63) related in a negative direction to her overall immigration stress and in a positive direction to her partner s supportive dyadic coping toward her (.45). No significant associations were found between the male overall immigration stress or supportive dyadic coping with either partner s relationship satisfaction. Immigration Stress Dimensions and Common Dyadic Coping. In additional analyses, seven separate models were estimated to examine the relationship between each immigration stress dimension and relationship satisfaction and whether common dyadic coping moderates these associations. Detailed results are presented in Table 4. For sake of clarity of presentation only an overview of the results is provided: Female relationship satisfaction was associated negatively with her own novelty (-1.12) and loss (-.33) dimensions and positively with the novelty dimension of the male immigration stress (.83). Common dyadic coping attenuated significantly the negative relationship of the female novelty (.96), not at home (.48), occupation (.88), discrimination (.48), and loss (.57) stress dimensions with her own relationship satisfaction. Common dyadic coping moderated in the same direction the negative association of female novelty (.53), language (.59), and loss (.44) with the male relationship satisfaction. Common dyadic coping also moderated the relationship between some dimensions of male immigration stress and each partner s relationship satisfaction but in the opposite direction: Common dyadic coping exacerbated the negative association of male novelty (males: -.70; females: -.87) and loss (males: -.38; females: -.32) with each partner s relationship satisfaction, of male language stress with his own relationship satisfaction (-.66), and of

21 IMMIGRATION STRESS 833 TABLE 5. Structural Models for Immigration Stress Dimensions with Supportive Dyadic Coping: Unstandardized Regression Coefficients, Variances of Dependent Variables and Regression Residuals, and Correlation of the Regression Residuals (N = 104 females and 104 males) FIS MIS FSDC MSDC FIS MSDC MIS FSDC Res. Var. Var. Res. Corr. Missing Family FRS MRS Novelty FRS MRS Not at Home FRS MRS Language FRS MRS Occupation FRS MRS Discrimination FRS MRS Loss FRS MRS Note. FRS = Female relationship satisfaction; MRS = Male relationship satisfaction; FIS = Female immigration stress; MIS= Male immigration stress; FSDC = Female supportive dyadic coping; MSDC = Male supportive dyadic coping. FIS/MIS FSDC/MSDC = interaction terms; Res. Var. = Residual (unexplained) variance of the dependent variable; Var = Variance (unconditional) of the dependent variable; Res. Cor. = Correlation of the residual terms of the dependent variables. Only unstandardized regression coefficients are presented since there is no statistical theory allowing for standardized regression coefficients with the LMS approach. ( ) indicates models with convergence problems, so results are presented without interaction terms. Parameters printed in bold type differ significantly from 0. (95% C.I.)

22 834 FALCONIER ET AL. the male occupation stress with the female relationship satisfaction (-.71). This set of negative effects still remained even when the interaction of female immigration stress and common dyadic coping was removed from each of the models. Overall, the interaction effects predicting female relationship satisfaction of common dyadic coping and female immigration stress dimensions of novelty, not at home, occupation, discrimination and loss are of comparable forms as depicted in Figure 3. However, for novelty, occupation, and loss also changes in male immigration stress should be considered. Immigration Stress Dimensions and Supportive Dyadic Coping. Table 5 presents the associations of each dimension of immigration stress, relationship satisfaction, and supportive dyadic coping as a moderator. For sake of clarity, only an overview of the results is provided. Female stress associated with missing family (-.76), novelty (-1.16), occupation (-.85), and loss (-.43) had a significant negative association with her own relationship satisfaction and there was no evidence of partner effects. None of the male dimensions of immigration stress had either actor or partner effects on relationship satisfaction. The male supportive dyadic coping moderated in a positive direction the negative relationship between the male and female relationship satisfaction with the female immigration stress associated with novelty (males:.88; females:.97), language (males:.62; females:.58), and loss (males: 0.46; females:.38). In addition, the male supportive dyadic coping ameliorated the negative association between the female stress related to missing family and discrimination and her own relationship satisfaction. The female supportive dyadic coping had no significant, independent association with either partner s relationship satisfaction and it even exacerbated the negative relationship of the male immigration stress associated with language (-.42) and his own relationship satisfaction and the negative association of the male stress associated with loss with each partner s relationship satisfaction (males: -.37; females: -.55). Comparing the interactions between dimensions of immigration stress and supportive dyadic coping to the interactions depicted in Figure 3, we conclude that the interactions of female missing family, novelty, language, discrimination, and loss with male supportive dyadic coping are of comparable form.

23 IMMIGRATION STRESS 835 DISCUSSION Findings from this study suggest that both Latino men and women experience comparable levels of relationship satisfaction and overall immigration stress. This is also true when stress associated with experiencing a novelty, discrimination, not feeling at home, language and occupational challenges, and missing family are considered and despite the fact that men have resided on average two years longer than women in the U.S. Gender differences in length of stay in the U.S. were not associated with significant gender differences (except for the sense of loss dimension of immigration stress) in immigration stress probably due to the fact that length of stay in the U.S. has only a modest relationship with either partner s overall immigration stress. This modest association also suggests that both partners in Latino immigrant couples continue to experience stress associated with their condition even years, and sometimes decades, after they have immigrated to the U.S. In fact, couples in our sample reported sometimes experiencing stress when they had been residing 13 years on average in the U.S. It is possible that immigrants experience stress differently over time, perhaps as anxiety and fear at first and more as frustration, sadness, and/or anger after some years. Nonetheless, length of stay in the U.S. was not associated with dyadic coping or relationship satisfaction, suggesting that the associations found between immigration stress, dyadic coping, and relationship satisfaction are not affected by the time that individuals have been living in the U.S. The main goal of this study was to investigate the association between Latino partners immigration stress and their relationship satisfaction and whether supportive and common dyadic could moderate such associations. As hypothesized, overall immigration stress was found to be negatively associated with relationship satisfaction but only for women. Latinas overall immigration stress was negatively related to their own relationship satisfaction but it did not seem to be associated with their male partner s. Within immigration stress, it was missing family, adjusting to new situations in the U.S., feeling a sense of loss, and experiencing occupational challenges that was associated with declines in Latinas relationship satisfaction whereas language and discrimination did not have such a relationship. It is possible that for Latinas those dimensions

24 836 FALCONIER ET AL. of immigration stress contribute to forming negative, pessimistic views about their lives and circumstances, including their couple s relationship. It is also likely that, similar to what happens with other sources of stress (e.g., economic stress; Falconier & Epstein, 2010), immigration stress affects their relationship satisfaction more indirectly by creating changes in the interactions with their partners. The immigration stress might increase hostile interaction with their partners, which in turn might create dissatisfaction with their couple s relationship. It is also possible that women dissatisfied with their couple s relationship might be more affected by immigration stressors. The cross-sectional nature of the data used in this study prevents us from ruling out this possibility. In contrast to women, Latino men s immigration stress did not seem to be negatively related to either partner s relationship satisfaction. This result suggests that Latino men may be coping with their own immigration stress in ways that do not seem to spill over into their couple s relationship. This possibility seems consistent with traditional gender role attitudes found among Latino immigrants in the U.S. (Falconier, 2013) by which women perform expressive functions and men are encouraged to keep their emotions and stress to themselves. In terms of the role of dyadic coping in moderating the relationship between immigration stress and relationship satisfaction, results from the present study suggest that common and supportive dyadic coping may be playing such a role but mostly regarding women s immigration stress. The couples conjoint efforts to cope with stress together were moderating the negative associations between almost all of the different dimensions of the female immigration stress and her relationship satisfaction. Similarly, although there is no main effect of either partner s immigration stress on the male s relationship satisfaction, the interaction of common dyadic coping and some of the female immigration stress dimensions (novelty, language barriers, and sense of loss) were positively related to the male relationship satisfaction. By contrast, the interaction of the male immigration stress with common dyadic coping was not associated with higher relationship satisfaction in either partner. Moreover, the interaction of the male immigration stress dimensions of novelty and loss with common dyadic coping was associated with declines in both partners relationship satisfaction. Similar to common dyadic coping, it depends on the partner s gender whether partner s support to cope with stress has a benefi-